Group Chat Interaction Engine

Problem It Solves

In group-chat scenarios, agents often reply with one-line acknowledgments instead of taking a clear position, explaining judgment, or participating in a structured discussion. This skill turns “workers” from single-shot responders into persona-driven participants: each worker has a persona card, and when a relevant topic appears, they speak from their own personality, expertise, and expression style rather than using a uniform customer-service tone.

How It Works

• Persona-driven: reads config/personas.json, which defines personality (cautious, practical, literary, rational, critical), trigger topics, speaking style, and reply templates.
• Topic triggering: when a group message matches keywords or a topic scope, the worker responds from the corresponding role perspective instead of giving a generic reply.
• Style control: output can be shaped into short and precise, code snippets, long analogies, numbered points, or direct critique, so different workers produce visibly distinct contributions.
• Runtime dependency: uses the Dashboard API to fetch context via GET chat and send discussion content via POST send.

Limitations

• Quality depends on the granularity of personas.json: overly broad triggers can make workers over-respond, while overly narrow triggers can leave them silent.
• Templates are scaffolds; the inserted topic context must be concrete enough to avoid vague slogans.
• The skill assumes access to the Dashboard API and the persona configuration file, so it is not suitable for fully offline, stateless, or externally disconnected deployments.